Reaching the Limits of Quantitative Life Cycle Assessment


Mark Rossi, PhD

Clean Production Action

June 2004


A Critical Review of

Life Cycle Assessment of PVC and of Principal Competing Materials

(A report commissioned by the European Commission,

authored by a consortium led by PE Europe GmbH, April 2004)






Clean Production Action (CPA) partners with environmental organizations, public health advocates, labor unions and community groups around the world to develop and build technical support for clean production policies. These policies and strategies promote the use of products that are safer and cleaner across their life cycle for consumers, workers, and communities





The report, Life Cycle Assessment of PVC and of Principal Competing Materials, summarizes the environmental problems and advantages of polyvinyl chloride (PVC) plastic across its life cycle, including its use in windows, pipes, wire and cable, flooring, and packaging.  The life cycle summary of PVC is based upon existing product life cycle assessments (LCAs) that are reviewed in the report.  The LCA review was commissioned by the European Commission and conducted by a consortium that was led by the German consultant group, PE Europe GmbH. 


Unfortunately the omission and selective reporting of data by the authors in their summary of the PVC life cycle along with the inherent limitations in the quantitative LCA method combine to create a flawed report.  Given these concerns, which are discussed in detail below, the conclusions emphasized and reached by the authors should be treated with suspicion. 


Readers should note that the European Commission, despite funding the study, has not embraced it and stamped the study with this disclaimer:


“The statements and conclusions expressed are those of the project consortium and should not necessarily be regarded as stating an official position of the European Commission” (p.2).



Failure to Acknowledge the Limitations of Life Cycle Assessment (LCA)


The authors of Life Cycle Assessment of PVC in their evaluation of LCA as an analytic tool fail to acknowledge many of the serious limitations to the use of LCA as a method for informing or making policy decisions.  They conclude that “with the development of the standardised method of LCA, LCA results can be used to obtain material decisions or choices with a ‘holistic’, ‘comprehensive’ and ‘objective’ life cycle view” (p.26).  Many have concluded otherwise, but the authors have chosen to ignore their concerns.


Arnold Tukker (1999), an LCA expert, has concluded that the underlying weaknesses of the LCA method are too great to withstand skeptical scrutiny:[1]


“Personally, I believe it will never be possible to solve controversial discussions about products with an LCIA [life cycle inventory assessment] method that is based solely on mathematical relations between interventions and protection areas.  There are simply too many uncertainties, there is too much ignorance, and they can only be overcome by all kinds of subjective, subtle, and basically value-laden choices. …


The authors of Life Cycle Assessment of PVC assume that values do not enter into the LCA method because it is constructed to be objective.  Hertwich, et al. (2000), however, concluded that “Due to the presence of dissimilar impact and the need to focus the analytical effort, preference values always enter the analysis, and so LCA results are not as value free as one might wish.”[2]  For Hertwich, et al. (2000), the presence of values is not a reason against LCAs, rather the presence of values means analysts need to clearly identify and state value-based choices.  The authors of Life Cycle Assessment of PVC did not identify nor state any value-based choices.


Furthermore, Stevels, et al. (1999) identified limitations with LCA that are especially relevant to the life cycle of PVC, including:


§      LCAs are “less effective where toxic/hazardous substances are involved.”

§      LCAs are “fairly cumbersome” in tackling recycling issues.

§      LCAs are a static, rather than dynamic analytic tool, therefore “taking the future into account is problematic, particularly for resources use.”


The difficulties of addressing toxic chemicals and recycling are highlighted below.  In addition, Stevels, et al. (1999), agreed with Tukker (1999) and concluded that “LCA is not yet appropriate for external comparison [i.e. public policy decisions].” [3]



Environmental Health Concerns of Very Toxic Chemicals are Downplayed


The concern raised by Stevels, et al (1999) on the weakness of LCAs in addressing toxic chemicals is affirmed by the summary of the LCA data presented in Life Cycle Assessment of PVC and the conclusions reached by the report’s authors. 


Example 1: Dioxins in Production Discounted


            Life Cycle Assessment of PVC:


“The occurrence of dioxins [during EDC/VCM production] is not included in the environmental impacts of LCAs” (p.53).  The authors conclude that this omission is justified because: “emissions of VCM, EDC, dioxins and other substances … will be close to or below the negligible risk level in modern facilities” (p.54). 


The report fails to note that the U.S. Environmental Protection Agency estimates, based upon industry data, that ethylene dichloride (EDC) / vinyl chloride monomer (VCM) production is among the top 15 sources of dioxin emissions in the United States. 


The study fails to note that under Article 5(c) of the Stockholm Convention on Persistent Organic Pollutants (POPs), ratifying nations agree to “Promote the development and, where it deems appropriate, require the use of substitute or modified materials, products and processes to prevent the formation and release of the chemicals listed in Annex C” [which includes dioxins]. 


Therefore, given that EDC/VCM production is a source of dioxins, that there is an international treaty calling for the elimination of dioxin sources, and that many EC nations (Austria, France, Germany, The Netherlands, and Sweden) have ratified the treaty, to state that dioxin emissions “will be close to or below the negligible risk level” is a misstatement at best and raises serious questions concerning the data analyzed to evaluate the life cycle impacts of PVC. 


The failure to address concerns with dioxins stems at least in part from the tendency of LCAs to only consider pollutants released in high quantities.  Since LCAs include emissions of chemicals such as nitrogen oxides (NOx) at the kilogram-level, they can only make sense of very small emissions if the chemicals are weighted.  For example, dioxin releases are measured in grams and dioxins can have adverse effects at the nano- to pico-gram level.  But dioxin releases in an LCA can only be meaningful if they are weighted, otherwise their releases are too small relative to other chemicals to be quantitatively meaningful.  Weighting the toxicity of chemicals is a complex process for which no widely agreed upon method has been developed. 


Example 2: Dioxins in Incineration Discounted


How the authors handle the issue of dioxin emissions from incineration is an example of selective citation or poor research.  They cite a study from 1989 which concluded that PVC contributes to dioxins from incinerators.  They then cite more recent studies to “conclude that the presence of PVC has no significant effect on the amount of dioxins” released from incinerators (p.81). 


Yet there are plenty of studies since 1989 that found a relationship between PVC and dioxin emissions from incineration.  For example, researchers working with full-scale combustion systems have found a positive relationship between the input of PVC and/or other chlorine sources and dioxin outputs in untreated and partially treated flue gas and/or fly ash.[4]  Katami et al. (2002) recently concluded that “PVC contributes significantly to the formation of PCDDs, PCDFs and coplanar PCBs from mixtures of plastics upon combustion.”[5]  And other researchers, while noting that incinerator operating conditions can be the primary influencing factor in dioxin formation, also confirm a clear role for PVC, among other contributors to chlorine content of waste, in the formation of elevated quantities of dioxins.[6]


How the authors handled the PVC-dioxin relationship in incinerators is either an example of selective citation of data or a case of poor research.  In either case their result is to provide data that dismisses the PVC-dioxin relationship and ignores recent studies that conclude otherwise. 


Example 3: Avoid or Barely Acknowledge Toxic Plasticizers


PVC requires the addition of softening agents -- called “plasticizers” -- to make it flexible.  However, these plasticizers are not bound to the polymer and leach out over time.  The authors acknowledge that:


§      PVC is a major consumer of phthalates (p.62) [PVC use accounts for around 90% of total global phthalate consumption]. 

§      DEHP is a category 2 chemical (p.62). [They do not state that a “category 2” chemical in the case of DEHP means it is a reproductive and developmental toxicant.]

§      Food is the main source of exposure to DEHP (p.61).

§      “Plasticisers can be released directly to air and to cleaning agents during cleaning” of vinyl flooring (p.79).

§      “Environmental contamination by this agent [the plasticizer DEHP] results from losses during manufacturing, packaging, storing and from flexible PVC articles” (p.61).


However, they do not acknowledge that:


§      DEHP is widely dispersed in the environment and that most (if not all) people in developed nations have phthalates in their body.

§      The EU Directive on General Product Safety (92/59/EEC) places an interim prohibition on the use of six phthalates (DEHP, DIDP, DINP, DBP, BBP, and DNOP) commonly found in soft PVC toys and childcare articles intended to be put into the mouth of children under the age of three years.

§      Patients receiving medical treatments that involve DEHP-containing PVC medical devices are exposed to the highest levels of DEHP of any segment of the population.

§      Alternative flexible plastics to PVC are inherently flexible and do not require the addition of plasticizers to achieve flexibility.


The case of plasticizers illustrates how the use of a product can result in multiple diffuse releases of small amounts of a chemical.  Each product in itself is literally a micro-smokestack, emitting small amounts of a chemical.  But taken as a whole the emissions from all these micro-smokestacks adds up with the result that DEHP is commonly found in tests of toxic chemicals in human bodies, and often it is found in the greatest concentration.  LCAs simply do not account for this level of complexity, assuming that each release of a product in use is just that: a single release rather than one release among many. 


The authors fail to acknowledge the cumulative results of many PVC products emitting small amounts of the same chemical nor do they wrestle with how, if at all, an LCA could address this issue.


Over-Emphasizing the Recycling Potential for PVC


Among the commodity plastics -- polyethylene, polypropylene, and PET -- PVC along with polystyrene are the least recycled plastics.  Yet the authors do their best to place a very positive spin on the opportunities for recycling PVC as evidenced by these “general conclusions” in the executive summary:


§      “Some new technologies exist, e.g. mechanical recycling based on selective dissolution, for recycling PVC in an economically feasible way” (p.12).

§      “In contrast to some metals, the recycling market of plastics, and therefore the demand in secondary material, is not yet established in an adequate way” (p.12). 


The first conclusion is highly unusual to highlight as a major conclusion (one of six general conclusions listed in the first page of the executive summary) in the context of an evaluation of LCAs because the focus of an LCA is on the state of products today not their future environmental potential.  Therefore, as part of reviewing the evidence on the PVC life cycle, the emphasis should be on the current state of PVC recycling -- which happens to be very low -- rather than on the possibility for future increases in PVC recycling.  While a brief mention of the new technology in the report is fine, it remains awkward in that every material that competes with PVC has some new element that will make it more environmentally preferable.  Why aren’t these new developments for other materials mentioned?  This is yet another example of selective reporting and emphasis in the report that favors PVC.


The report does state that overall post-consumer recycling for PVC in Europe is 3%, with very little of that closed loop recycling, and the vast majority being “mixed plastic recycling” that results in material of “low commercial value” (p.83).  However, this important statement of reality is buried on page 83 and excluded from the executive summary.


Furthermore, the very low PVC recycling rates means recycling has a negligible effect on an LCA outcome for PVC because the opportunities for offsetting the impacts of virgin production by using recycled content are essentially not realized.  When the recycling stage of a product’s life cycle is incorporated into a LCA it will reduce the emissions from virgin production (assuming the product contains recycled content) or will potentially reduce emissions from disposal if the product is down-cycled -- recycled into a different product with lesser technical performance needs for the material.  More relevant would have been an analysis of why PVC recycling rates are so low.


The second general conclusion implies that the recycling rates for all plastics are similar, when that is not the case.  Had the report evaluated plastics recycling rates and markets it would have found PVC to have among the lowest recycling rates for a commodity plastic.


Furthermore neutral analyses of the recyclability of plastics have concluded that PVC is among the most difficult to recycle plastics.  The difficulty of recycling PVC, especially closed-loop recycling, has led the automobile industry, for example, to target PVC for elimination.  Driven by end-of-life vehicle directives in Europe and Japan to increase recycling rates for automobiles, automakers are evaluating their use of plastics and selecting plastics that can be recycled back into the same product.  The European automaker, Opel, for example, classified plastics according to their recyclability.  PVC was next to last on the list in terms of recyclability, with PVC only more recyclable than a “mixture of incompatible products” (see Table 1 below).  All of the automakers have reached the same conclusion as Opel: that polypropylene and polyethylene are the easiest to recycle plastics and PVC is among the most difficult.



Table 1.  Opel Priority List for Plastics with regard to Recycling Aspects

Avoid -----Increasing Priority------Prefer

Polypropylene, Polyethylene

Polyoxymethylene (POM), Polyamide, Thermoplastic Urethane (TPU)

Acrylonitrile Butadiene Styrene (ABS), Polymethylmethacrylate, Styrene Maleic Anhydride (SMA) copolymer, Acrylonitrile Styrene Acrylate (ASA), Styrene Acrylonitrile (SAN)

Polycarbonate, Polyethylene Terephthalate (PET), Polybutylene Terephthalate (PBT)

Thermoplastic Elastomer (TPE)


Sheet Molding Compound (SMC)


Polyvinyl Chloride (PVC)

Mixture of incompatible materials

                        Source: Opel, Environmental Report 2000/2001.




The authors treat the recycling phase of PVC’s life cycle in a manner that emphasizes opportunities and downplays problems.  A balanced report would have emphasized the current difficulties of recycling PVC, the neutral evaluations of the recyclability of PVC, as well as the possibilities for increasing PVC recycling rates. 


Material versus Product Evaluations


The authors state that LCAs are most relevant at the product level because it enables analysts to evaluate the comparative benefits and problems of products across their entire life cycle: resource extraction, production, use, and recycling/disposal.  They state that material level assessments are insufficient because the analysis only covers the extraction and production stages, missing the use and recycling/disposal stages (p.12).  Their conclusion that only a product level analysis is appropriate is based upon the current limited design of the LCA method. 


The LCA method is designed to evaluate products, not materials, across all life cycle stages.  That does not mean, however, that a life cycle evaluation covering all stages of a material is infeasible.  Analyses can be done for materials across all life cycle stages (extraction, production, use, disposal/recycling) by evaluating the contribution of the materials to specific environmental problems.  For example, we can evaluate and compare the contributions of specific materials to the formation of persistent organic pollutants (POPs) across the entire life cycle of the materials.  Such an assessment would need to factor in contributions from different representative products made from the material to reach conclusions about the material in terms of specific endpoints.  While a material life cycle assessment will look different than a product LCA, it is still quite feasible.  What is necessary is a clear definition of scope, endpoints to be examined, and methods for evaluating endpoint impacts.  Unlike a product LCA, a material approach begins from a set of pre-determined endpoints for analysis rather than trying to evaluate all potentially relevant endpoints.  This is a value decision that analysts must make and must be explicit about making.


A materials life cycle assessment method is needed to address a failure of product LCAs: the failure to account for the cumulative life cycle problems that arise over the use of a material in multiple products.  As highlighted below, this is clearly the case with phthalates leaching from flexible PVC products as well as with dioxin emissions from the manufacture and incineration of PVC.  While each individual PVC product’s contribution to DEHP or dioxin emissions will be small, cumulatively they are significant.  This is especially true for persistent and bioaccumulative toxics like dioxins, which diffuse widely throughout the environment and accumulate over time building up to hazardous concentrations in our bodies.  Therefore LCAs are necessary that account for the cumulative effects of using a material in multiple products across the entire life cycle of the material/product (not just the extraction and production stages).


By defining product LCAs as the only approach to evaluating life cycle concerns, the authors are defining away the problem of the cumulative impact of multiple small exposures by multiple products.  By design the tool will miss the problem.





Through a combination of omitting key data and policy decisions, engaging in the selective citation of studies (or poorly conducting research), and relying upon a method -- quantitative life cycle assessment (LCA) -- unsuited for the task, the result is a study whose conclusions must be read with a healthy dose of skepticism.  Given these problems it is impossible to know the validity and relevance of the conclusions reached by the authors.  We conclude from this review that the study -- Life Cycle Assessment of PVC and of Principal Competing Materials -- is fundamentally flawed and that it should not be used or considered as a relevant comparative analysis of the life cycle problems associated with PVC and its competing materials.



[1] Tukker, A., Life Cycle Impact Assessment – Some Remarks.  Life Cycle Impact Assessment of SETAC-Europe (Second Working Group – WIA-2).

[2] 2 Hertwich, EG, JK Hammitt, WS Pease.  2000.  “A Theoretical Foundation for Life-Cycle Assessment.”  Journal of Industrial Ecology, v.4, n.1: 13-28.

[3] Stevels, A., H. Brezet, J Rombouts.  1999.  “Application of LCA in Eco-Design: A Critical Review.”  The Journal of Sustainable Product Design, April: 20-26.

[4] For example, see:

-- Oberg, T., Ohrstrom, T., 2003.  Chlorinated aromatics from combustion:  Influence of chlorine, combustion conditions and catalytic activity.  Environ. Sci. Technol. 37: 3995-4000.

-- Hunsinger, H., Jay, K., Vehlow, J., 2000. Formation and destruction of PCDD/F inside a grate furnace. Organohalogen Cpds. 46:86-89.

-- Yamamura, K., Ikeguchi, T., Uehara, H., 1999. Study on the emission of dioxins from various industrial wastes incinerators. Organohalogen Cpds. 41: 287-292.

-- Tagashira, K., Torii, I., Myouyou, K., Takeda, K., Mizuko, T., Takushita, Y., 1999.  Combustion characteristics and dioxin behavior of waste fired CFB. Chemical Engineering Science 54:5599-5607.

--  Costner, P.,1998. Correlation of chlorine input and PCDD/PCDF emissions at a full-scale hazardous waste incinerator. Organohalogen Cpds. 36: 147 – 152.

-- Tuppurainen, K., Halonen, I., Ruokojarvi, P., Tarhanen, J., Ruuskanen, J., 1998.  Formation of PCDDs and PCDFs in municipal waste incineration and its inhibition mechanisms:  A review.  Chemosphere 36: 1493-1511.

-- Manninen, H., Peltola, K., Ruuskanen, J., 1997. Co-combustion of refuse-derived and packaging-derived fuels (RDF and PDF) with conventional fuels.

 Waste Manage. Res.15:137-147.

-- Manninen, H., Perkio, A., Vartiainen, T., Ruuskanen, J., 1996. Formation of PCDD/PCDF: Effect of fuel and fly ash composition on the formation of PCDD/PCDF in the cocombustion of refuse-derived and packaging-derived fuels. Environ. Sci. & Pollut. Res. 3 (3): 129-134.

-- Wanke, T., Vehlow, J., 1996. The influence of flame retarded plastic foams upon the formation of Br containing dibenzo-p-dioxins and dibenzofurans in a MSWI. Organohalogen Compounds 28: 530-535.

-- Huotari, J., Vesterinen, R., 1996. PCDD/F emissions from co-combustion of RDF with peat, wood waste, and coal in FBC boilers. Haz. Waste & Haz. Materials 13(1): 1-9.

-- Vesterinen, R., Flyktman, M., 1996. Organic emissions from co-combustion of RDF with wood chips and milled peat in a bubbling fluidized bed boiler. Chemosphere 32(4): 681-689.

-- Moller, S., Larsen, J., Jelnes, J.E., Faergemann, H., Ottosen, L.M., Knudsen, F.E., 1995. "Environmental Aspects of PVC." Environmental Project No. 313. Denmark: Ministry of the Environment, Danish Environmental Protection Agency, 1995.

-- Thomas, V.M., Spiro, T.G., 1995. An estimation of dioxin emissions in the United States. Toxicol. and Environ. Chemistry 50:1-37.

[5] Katami, T., Yasuhara, A., Okuda, T. and Shibamoto, T. (2002) Formation of

PCDDs, PCDFs and coplanar PCBs from polyvinyl chloride during combustion in an incinerator. Environmental Science and Technology 36(6): 1320-1324.

[6] Hatanaka, T., Imagawa, T. Takeuchi, M. (2000) Formation of PCDD/Fs in artificial solid waste incineration in a laboratory-scale fluidised-bed reactor: Influence of contents and forms of chlorine sources in high temperature combustion. Environmental Science and Technology 34(18): 3920-3924

Wikstrom, E., & Marklund, S. (2001) The influence of level and source on the formation of mono- to octa-chlorinated dibenzo-p-dioxins, dibenzofurans and coplanar polychlorinated biphenyls during combustion of an artificial municipal waste. Chemosphere 43(2): 227-234

Yasuhara, A., Katami, T., Okuda, T., Ohno, N., & Shibamoto, T. (2001)

Formation of dioxins during the combustion of newspapers in the presence of sodium chloride and poly(vinyl chloride). Environmental Science and Technology 35(7): 1373-1378